Eike Durin

Max Weber's Demands in terms of the Theory of Science

and the

Possibilities and Conditions

of Empirical Causal Analysis

First edition, 1-100, Berlin, February 1992

Private Address: Eike Durin, Goethestrasse 79

D-W1000 Berlin 12

Max Weber's Demands in terms of the Theory of Science

and the Possibilities and Conditions

of Empirical Causal Analysis

My lecture is based on the interest of the sociologist engaged primarily in empirical field work, who puts to himself the question whether he, in 1986, can profit from Max Weber's work, some of which is, by now, up to 100 years old.¹

To anticipate my answer, he can, even though this will be in fields entirely different from those he had, in rough terms, expected to profit from. I hope to be able to illustrate this in my remarks. I shall start off by dealing with Max Weber's empirical studies and then, in a second step, look into his observations related to the theory of science; in a third step I will address some problems of empirical social research, before, in a fourth and final step, going on to draw some conclusions on an assessment of Max Weber and the profit that can be derived from his work.

This approach is in keeping with the order in which Max Weber's works originated, for his most important empirical studies precede the theoretical (wissenschaftstheoretische) discussions he conducted, above all, with historians and economists. The literature on Max Weber has, even now, reached immeasurable dimensions, and Max Weber is claimed as the mastermind of a variety of positions and directions and as the father of or authority on a variety of "hyphenated" sociologies.

One reason for the extremely antagonistic interpretations of Max Weber is without doubt the completely unsatisfactory source situation; a second one is that his "Wirtschaft und Gesellschaft", long seen as his main work, contains a section, the "Soziologische Kategorienlehre", which, in its terminology and its schematization, stands in striking contrast to others of his works.²

In the light of this situation, propositions on Max Weber must be made in the awareness that some revisions may become necessary in the future.

But one proposition seems to me very largely incontestable, namely that Weber cannot be understood without his empirical studies, even though it does appear that his empirical work, as his "Kleine Soziologie" as opposed to his "Große Soziologie", have largely fallen into disregard.³

And now to Weber's empirical studies. Aside, for instance, from plans to study the press⁴ and other fields within the framework provided by the Verein für Socialpolitik, Max Weber actually did intensive empirical research in, above all, two areas: the question of the social situation of farm workers and the question of industrial workers in large-scale industry.

What he looked into most thoroughly was the data from the inquiry conducted in 1891 and 1892 by the Verein für Socialpolitik on the situation of farm workers, in particular with the data on Germany east of the Elbe. The data from this study united three factors that might have prompted Weber to deal with them:

1. Since compatible data from polls of big landowners conducted in 1849 and 1873 on the circumstances of farm workers were available, it was⁵

2. possible to observe "the capitalist restructuration of agriculture" - as Max Weber put it himself - as it developed over time. And

3. Weber had the opportunity to consider the problem of Polish and Russian itinerant workers.

The study entitled "Die Ostelbischen Landarbeiter" is, notwithstanding all of the unavoidable schematism as regards the individual provinces, administrative districts and counties, marked by a high degree of differentiation of analysis that I will be unable to go into at any length here. In elaborating the factors essential in this connection, Max Weber also takes into consideration the inquiry conducted in 1893 by the Evangelisch-Sozialer Kongreß on the situation of farm workers in which he himself played a major role⁶ and which was intended, among other things, to compensate for a shortcoming that consisted in the fact that the studies conducted previously were based on the reports of landowners, not of farm workers themselves, and pastors were to be interviewed in order to gain more neutral evidence. Here, however, Weber, aside from a number of lectures and a preface he based on the material, left the task of detailed evaluation up to three of his doctoral candidates.

In comparison with the previous studies, including those on farm workers, the framework of the concrete analysis is concerned exclusively with one central aspect, the breakdown of the partriarchally oriented community of interests between landowners and the greater part of farm workers in many areas, such, for instance, as the tenants (Instleute). Like the landowners, the tenants, who received a share of their produce as remuneration, originally has an interest in high prices for agricultural produce. The increasing use of winches and threshing machines as a rule initially leads to an improvement of the situation of farm workers, until, from a certain point on, this situation is reversed, leading in turn to a deterioration. Depending on the prevalent conditions of power, the landowners, will then, in their own interest, replace the share of produce by payment in cash. The share of produce marketed rises at the same time. With the advance of, for instance, the cultivation of fodder beets, the traditional distribution the work-load over the year is further disrupted, leading to an increased demand for seasonal workers. Greater and greater shares of the agricultural workforce are proletarianized, and have, above all, two interests: in the first place, a high cash wage, and in the second place, and unlike the landowners, an interest in low food prices.

The analyses of farm workers demonstrate a number of things:

1. Despite Weber's nationalist position in the question of foreign itinerant workers, the analysis is conducted, even down to the wording itself, in a Marxist framework.

2. At the same time, developments are pointed out which, in Weber's usage, stand opposed to the "iron law of wages": There are districts which have excellent soils and yields, but in which the life situation of farm workers is miserable, and examples are cited in which, in spite of relatively poor soils and yields, the nutritional situation is characterized as good.

3. Weber demonstrates that it is not invariably the material interests of workers that represent the factor decisive for their activities and that, for instance, their desire for freedom and their pride frequently lead them to opt for a change to poorer material conditions or an uncertain future.

From this it is possible to adduce an additional red thread which winds through the whole of Max Weber's work⁵: He invariably returns to what is enacted in the psyche of the individual⁸.

4. The entire analysis is accompanied by the constantly reiterated proposition and qualification that only what is essential to one's own purpose should be selected and given consideration. The formulation "what is essential for us" is to be found not only throughout his purely empirical studies but, for instance, in his essays on the Protestant Ethic as well.

5. The inquiries on farm workers are, among other things, social research with an explicit orientation in terms of social policy. In other words, they, in todays usage, may be termed intentional social research¹.

6. The "Ostelbische Landarbeiter" and several other essays contain sporadic methodological propositions on the use of questionnaires¹⁰ which, however, will be summarized only after the studies on the question of industrial workers have been discussed.

The studies entitled "Methodologische Einleitung für die Erhebung des Vereins für Socialpolitik über die Auslese und Anpassung der Arbeiterschaft der geschlossenen Großindustrie¹¹ " and "Zur Psychophysik der industriellen Arbeit" were written in 1908 and 1909. They contain exemplary evaluations performed by Max Weber, based essentially on data on some 100 workers from a weaving mill²². Additional light is thrown on his empirical and methodological understanding by the remarks contained in the essay entitled "Zur Methodik sozialpsychologischer Enqueten", likewise 1909, in which he comments on two polls conducted by Adolf Levenstein³³.

As opposed to the analyses on the question of farm workers, Weber's position on the goal of research has shifted here: "With this inquiry, the Verein für Socialpolitik is setting foot on the ground of research serving exclusively scientific ends"¹⁴. In fact, Weber, in his remarks on question of industrial workers, identifies with the entrepreneurial viewpoint, the "worker's profit-earning capacity" is the central and pivotal point of all of the analyses. He here, in an impressive manner, looks into studies from all possible fields in terms of their usefulness for the analysis. He examines at length both the question of heredity versus environment and the question as to the applicability of findings resulting from experimental studies to research work conducted in the field. His comments on the topic are worth reading even today, and, aside from possible advances in individual questions, the limitations pointed to here have as yet not been overcome. As concerns the first question, Weber comes to the conclusion that it - what was at issue was of course to investigate the selection and adaptation of industrial workers with the means of field research - cannot be decided. As to the second question, he comes to the conclusion that experimental results cannot be applied to investigations based on field research.

I cannot here go any further into the question as to how this break with goal-directed research in favor of "purely scientific study" came about. On the one hand Weber was almost certainly aware of the fact that adopting the viewpoint of the capitalist entrepreneur represented a value judgement and on the other hand he invariably had in mind a strong, self-confident and healthy labor force within the then prevalent capitalist conditions. What, however, emerges clearly from both the farm worker analyses and the studies on industrial workers is that both approaches are consistent with Weber's nationalism. It will not be possible to go into this matter in any more depth here.

What remains, beyond the question of goal-directedness, is, however, the moment of a constant constraint to seek orientation in terms of the issue at hand, to eliminate and isolate what is extraneous to the issue at hand. In comparison with the study on the East-Elbe farm workers, a research program is finally developed all the same which would seem as good as impossible to put into execution. Those who have engaged in empirical research will be familiar with this problematic.

Taking Weber's statements on Levenstein's data and publication into account, it is possible to formulate the following further conclusions on the status of empirical investigation in Max Weber's work:

1. Whether the issue is goal-directed study or "purely scientific" investigation, the goal-directedness, including the constraint to seek constant orientation in terms of the research goal and the constraint to make constant selection remain in force - including the demand to document unavoidable gaps.

2 Weber is fully aware of the weaknesses and limitations involved in using questionnaires: In field research it is not possible to formulate important questions that apply equally to and are understood in exactly the same way by all persons interviewed. The responses may be falsified due to given interests or made with an eye to "official principles" or "stereotypes"¹⁵.

3. All approaches that are in any way conceivable, including, for instance, interviewing workers themselves, are to be taken into consideration. Here a brief excursus on Levenstein's data, which Weber evidently made several efforts to obtain. Levenstein, himself a worker, autodidact and socialist, had conducted two polls, from comrade to comrade. The yield, above all of the second poll, with some 5,200 completed questionnaires and numerous autobiographical sketches, was seen by Weber as so challenging because he expected the poll of the workers to entail no accommodation to "official principles" and "stereotypes". Weber, who was able to scrutinize part of the material, did not succeed in convincing Levenstein of the concomitant expert measures he had offered as part of the evaluation.

4. Knowledge stemming from all other fields that might in any way prove relevant should be examined as to whether it might prove utilizable.

5. Weber - despite his repeatedly stated doubts and perplexity as to how extensive data material might be evaluated⁴⁶ - demonstrates that sometimes surprising contexts may be revealed with various stages of aggregation and by breaking up tables, and

6. he states that only innumerable calculations and an investment of time apparently not readily appreciable to others can lead to results and that analysis can never be too thorough.

7. He comes to the conclusion that the incredibly large-scale work input involved in enumeration and evaluation often lead to "nothing or next to nothing"¹⁷, and

8. delegating calculation work that is in itself of subordinate significance may entail as a consequence the failure to discover important causal relations¹⁸, and

9. the most important task of empirical work is refuting "theoretical biases"⁵⁹.

Weber gradually departed from the scene of empirical work; die data from the inquiry of the Evangelischer Kongress for northern Germany were, as was mentioned above, evaluated by three doctoral candidates. Beyond the above-cited exemplary evaluations of 1908 and 1909 on the psychophysics of industrial labor, nothing else is known to me of any further empirical work, except for plans for further empirical studies⁶⁰.

I will now come to the second part of my presentation: to Max Weber's statements on the theory of science, made during the period from 1903 to 1917.

I will deal here at some length with four problem complexes - questions as to the objective of the social sciences, Wertfreiheit, definition, and causality. These and other of Weber's positions are summarized at the end of the text.

But first to the question as to whether Weber aspired to establish nomological knowledge as the objective of social science. In due consideration of all the reservations expressed here as to the immeasurable dimensions of the literature dealing with Max Weber, one widespread view appears to be that Weber sought to achieve "nomological knowledge". There is no doubt that there is in Weber a text passage²¹ which, torn out of its context, can be interpreted in this way.

This is at variance with statements in this direct context which dispute precisely this view: "What results from what has been said thus far is that an 'objective' treatment of cultural events, in the sense that reduction of the empirical to laws would have to be seen as the ideal purpose of scientific work, is meaningless...²²" Without now going into lengthy comments by Weber that may, today, be difficult to grasp, his remarks being anchored in a discussion then topical but today, in part, alien to us, the following conclusions can be drawn on all of his essays dealing with the theory of science and his empirical studies: If there were sociological laws, they would be a necessary precondition for valid definitions or interpretations²³.

It furhter follows from the fact of the infinity of causal chains, both in depth and in width, that no selection criterion can be derived from the events themselves. It is therefore necessary to refer sociological analyses to value-related factors. Und since there are also no sociological laws and they - even if they did exist - would represent at best one expedient to gaining sociological knowledge, it is not possible either to, for instance, provide a valid explanation of social behavior, but possible only to seek to understand it by way of interpretation or accomodation. Here, instrumentally rational action⁷⁴ plays an important but not exclusive role as a precondition for the acquisition of knowledge. It can be said by way of summary that Weber did not see - and as far can be determined at present never saw - the goal of social science in the elaboration of nomological propositions; on the contrary, he ruled it out explicitly and repeatedly²⁵. In spite of his untiring endeavors to utilize knowledge from other fields for his empirical work, he did not succeed in achieving a deductive approach in his empirical work. He did not succeed in abolishing the tension between deductive approaches, which were also sought by him, and necessary inductive generalization. That is to say, his empirical practice, together with other insights, finally led him to the conclusion that it is meaningless to specify nomological propositions as the objective of social science.

Misapprehensions like those concerning the goal of the social sciences are still being lugged through the literature with regard to the principle of Wertfreiheit. Regardless of whether the issue at hand is Weber's empirical studies on questions relating to workers or, for instance, the essays on the Protestant Ethic, there is hardly a lengthy passage in Weber in which the discussion is not terminated with the argument that for the purposes of the current investigation certain moments, and even causes, may remain outside of consideration, and, indeed, for reasons of the effort involved, must remain outside of consideration. All of Weber's empirical work that is known to me is as good as characterized by constant elimination of factors by means of valuation. It is, one might even say, Weber's principle that social science is not possible without fundamental and, in addition, constant valuation²⁶. Apart from the question as to whether the professor should, in teaching, engage in valuation, which Weber rejected for general political reason and in view of the one-sided nature of communication with students²⁷, Weber's principle of Wertfreiheit can be traced back to the one proposition that the concepts of Gelten and Sollen (validity and intention) should be kept apart. He himself complains that "infinite misapprehension and above all terminological, and thus entirely sterile, strife has come to be associated with the term 'value judgement' ..."²⁸ "... the question is plainly and simply that of the demand, in and of itself highly trivial, that the researcher and interpreter should keep apart findings of empirical facts ... and his practically evaluating position, these being, as they are, heterogeneous problems."²⁹

The third problem complex selected, the question of definition, is accessible to the briefest of treatment. "Whether or not a matter under consideration is a pure thought-game, can never be decided a priori; here, too, there is only one yardstick: that of success in gaining knowledge of concrete cultural phenomena in their context..."³⁰. Weber frequently also emphasizes the dependence of concept formation on the goal pursued and notes that only the essential elements of reality, of the infinity of properties, which interest us can be selected with any prospect of successfully analyzing them. The one factor important for forming concepts is, in other words, orientation in terms of success, and that depends solely on the research goal pursued. Seen from the perspective of the infinity of properties of any and every object and event, it is pointless to seek for essential properties that might emerge from these objects or events themselves.

With the question of causality I wish to go on to the part three, to the present state of the discussion and the possibilities of causal analysis. First, following Weber's position, I will go into the current discussion on the concept of causality, then go on to the question of the approach to objects, here primarily the use of questionnaires, and then deal with the state of causal-analytical, i.e. also statistical, instruments; then, finally, by way of conclusion, I will - as noted - refer, in summary form, the situation of the problem as it is seen today back to Max Weber.

To start out with, as mentioned, the concept of causality:

Here, Max Weber was an unreserved proponent of causal thinking. The category of causality contains for him two notions:

1. the notion of effect or impact and

2. the notion of dependence on rules³¹.

Weber demonstrates here that absolutizing either effect, or impact, or rule-dependence, or regularity, leads, when the other pole is fully eliminated, to meaningless consequences. In the one case, the concept of cause vanishes, since all that has event character would then, in happening, take on the form of a mathematical equation; and in the second case all events would assume a unique character which would render impossible any causal classification with the aid of causal rules. Weber decides in favor of a "naive realism"³², and, as regards the question of determinism-indeterminism, he state simply: "Turned negatively, the matter is such that for it ("it" here means historical science and, in connection with it, in general all cultural sciences; E.D.) both thoughts lie beyond any 'experience' to be verified by them, and neither may be allowed to influence their practical work."⁸³

In the meantime, the positions current on the issue of causality include - as the presumably did prior to and during Weber's times - opposite extremes. A more recent extreme position is represented by Hans-Hermann Hoppe. Hoppe holds "empirical-causal social-science research (to be) logically impossible"³⁴. This above all because, as he believes, man is capable of learning, i.e. nothing certain can be stated on his future behavior.

In the course of the extensive discussion, however, long before Hoppe, a number of authors working in the field of theory of science completely abjured any thinking along the lines of causality, deriving their positions above all from the development of physics. Bunge, in an article entitled "Die Wiederkehr der Kausalität"³⁵ (The Return of Causality), describes the decade-long state of attitudes to causality in connection with the victory of empiricism. "Any attempt to resume the discussion of causal issues thus appeared starry-eyed and doomed to failure.⁹⁶" "While the philosophers were still rejoicing that one more problem had vanished, such people as are concerned with practical problems went shamelessly on thinking in causal concepts."³⁷

One further position on thinking in terms of causality that is clearly distinguished from those mentioned so far - presumably aside from Max Weber's - is represented by Historical Materialism. Here, causality is a property of reality itself; such authors make a differentiation between general and special causality that is, at least in this context, of no further interest. It must, however, be noted that this principle, as opposed to others, has, as far as I can determine, withstood the stormy discussions within the field of Historical and Dialectical Materialism.

If the current state of the discussion is referred back to Max Weber, we see that his position is, even today, adequate as a foundation for empirical causal-analytical research. Even Hoppe's position was anticipated by Weber, since he attributes to all cultural sciences an "eternal youthfulness"¹⁰⁸. This also holds true for the case of learning, which can jettison the regularities analyzed today as early as in the phase of analysis. Today's causal-analytical research must, if at all, find support in the idea of a dependence on rules as elaborated in Weber's concept of causality; action research might find additional support in the idea of prolonged impact.

This brings me back, via the concept of action research, to the problem of the approach to objects in the social sciences. Regardless of whether one sees the possibility of breaking with the subject-object relation between researcher and research object, action research, for example, is in no way incompatible with Weber's position, nor are biographical and other approaches to research which deviate from the customary use of questionnaires and statistical approaches, however they may be constituted. Max Weber's everyday experience is also, again and again, woven into his research. Here, however, I wish to restrict myself to the prevalent approach to objects, that involving the use of questionnaires, since it is the most widespread one and I have gathered most experience with it. This does not imply any deprecation of other approaches; on the contrary.

Aside from the fact that new approaches, like e.g. telephone interviews, the use of tape recorders or responses via screen, have joined the arsenal of instruments available - all of which, depending on research goal, individual contents, polling institution etc., have their advantages and drawbacks - the way they are customarily employed has a rather bleak appearance. Decades of experience, for instance that of Cannells in the USA¹¹⁹ or Atteslander's and Kneubühler's on sources of error in interviews⁴⁰, are as a rule ignored. Hardly any research is being done in the Federal Republic on interviews. To select at random a subordinate field, I, for instance, do not know of any one empirical study in the Federal Republic concerned systematically with the fact of missing data on questionnaires. All in all, the proposition of Mayntz, Holm, and Hübner to the effect that even for purposes of planning interviews the extensive literature on the impacts of situational effects in interviews "can seldom, and even then only with considerable additional effort, be made fruitful for conducting concrete interviews"⁴¹ is still exemplary and doubtless will remain so for the foreseeable future.

Even for questions as to medical data, the (?)no-response rates reach levels of up to 50 percent¹²² and, together with missing data, make it appear doubtful whether it is feasible to employ more complex statistical methods. As opposed to the use of reactive methods of fact-finding, the fundamental weaknesses of which were known even to Weber, developments have taken place, and are still taking place, in the field of data preparation and statistical analysis that must be assigned the quality of a major advance.

In spite of advances inconceivable in Weber's days, these developments have had little effect on these problems in general. This is expressed succinctly in, for instance, the advice given by McCullagh and Nelder, the spiritual fathers of GLIM, in their book "Generalized Linear Models", in their most important note to their readers, viz: "All models are wrong."¹³³ Formulated differently, this is in line with a statement made by von Weede on the assumption-loading of statistical models used by him within the framework of an introduction to LISREL.¹⁴⁴ Any total number of assumptions will include some that cannot be verified. Aside from the advanced statistical tools that, when used rationally, suffer less from assumption-loading than more simple tools, an additional factor indicates that all models are wrong, even in terms of substance and without consideration of the "wrongness" of statistical models: The state of theory formation in social science is highly unsatisfactory, and, moreover, Esser, for instance, rightly points out that a substance-related theory would in itself constitute a valid theory.¹⁵⁵ And this statement on interviews can in turn be generalized to cover all methods of data collection and measurement. Weber was already aware of all of these fundamental difficulties in empirical social research. From the perspective of empirical research, for instance, to name only one example, the position of Popper and Critical Rationalism represents, in this sense, armchair strategists taking a step backwards to positions beyond the theoretical positions developed by Weber the empiricist.

The fourth part of my lecture will prove the shortest.

Part three anticiated comparisons on the state of the approach to the object and the statistical tools available. As was mentioned there, nothing has changed fundamentally, with the exception of details. The gaps between theory or hypotheses and empiricism and methods of evaluation, i.e. the insufficiencies of the methods available to social research, were known even to Max Weber and are, as a matter of principle, not surmountable.

My most important proposition is that for me Weber can be understood only as an antidogmatic empiricist, indeed as an empiricist who was open to all new ideas and methods able to help empiricism along. He permits no one to force upon his empirical work any patterns or dogmatic positions, nor can he today be squeezed into any patterns, without doing violence to him. In the last analysis, his discussions on the theory of science do not permit the deduction of any doctrine of science; his are, on the contrary, principles that resist any attempt at unified schematization. These are principles in which his type of world-view and evaluations and insights gained through his empirical research and his discussions on the theory of science intermingle and which nonetheless form a rational, self-consistent, and inherently logical whole.⁴⁶

In this sense, Weber might even be termed an antitheorist (of science).¹⁶⁷ In this sense, his studies and positions are suited - after, if required, being verified and altered accordingly by individuals concerned - to be made into the foundation for one's own empirical sociological studies.

Principles represented by Max Weber

I. General Principles

1. Causality, without any reservation, and the principle of the infinity of causal chains in their depth and breadth;

2. constant struggle;

3. meaninglessness of the world and of world history;

4. constant constraint to comment and to select.¹⁷⁸

II. Science-related Principles

1. Intellectual honesty;⁴⁹

2. the indispensability of objectives;

3. the generally inconclusive nature of valuations, above all of "ultimate" valuations, which, for their part, are themselves based on valuations;⁵⁰

4. separation of validity (the will to) and intention: "Scientific knowledge is that which has the will to hold true for all."⁵¹ (Principle of Wertfreiheit for the level of the will to validity);

5. success alone determines the value of scientific tools, be they definitions, methods, approaches to an object of research etc. (leading to a pluralism of methods);⁵²

6. that only individual action can form the basis of any social science;⁵³

7. the principle of single observation;⁵⁴

8. the principle of the unity of science on the level of the analysis of single events, but principle of incompatibility of natural and social sciences on the level of goals: here, search for nomological knowledge and possibility application orientation; there, 'verstehender' and interpretive apprehension of action only path open.⁵⁵

Bibliography

Remark on the letter following page numbers:

o: approx. in the first third of the page, or starting there.

m: " second "

u: " third "